Comprehensive evaluation of multisource aerosol optical depth gridded products over China. (1st June 2022)

• Record Type: Journal Article
• Authors: Jiang, Daoyang; Wang, Lunche; Yi, Xiuping; Su, Xin; Zhang, Ming
• Is Part Of: Atmospheric environment. Volume 278(2022)
• Publication Date: 2022-06-01
• Subjects: AOD -- Satellite products -- DB algorithm -- Bilinear interpolation algorithm; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2022.119088
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: Aerosol optical depth (AOD) is one of the most crucial parameters for reflecting aerosol characteristics. This study systematically evaluated daily AOD from Modern-Era Retrospective Analysis for Research and Applications Version 2 (MERRA-2), Visible Infrared Imaging Radiometer Suite (VIIRS), and Moderate-resolution Imaging Spectroradiometer (MODIS) daily gridded products over China from 2012 to 2019. Both the VIIRS aerosol Deep Blue (DB) and MODIS Deep Blue algorithm products had significant missing values in areas, for example, the Qinghai-Tibet Plateau, Xinjiang, and Northeast China. High AODs were mainly concentrated in areas closely related to economic development and population density geomorphology factors, such as the North China Plain, Yangtze River Delta, Pearl River Delta, and Sichuan Basin. Notably, VIIRS products captured higher AOD values in desert regions than the other products. In overall accuracy, MODIS DB product was characterized by a correlation coefficient (R) of 0.82, root mean square error (RMSE) of 0.21, and mean absolute error (MAE) of 0.14, and VIIRS (RMSE = 0.23 and MAE = 0.15) and MERRA-2 (RMSE = 0.27 and MAE = 0.17) had a slightly larger bias. MODIS performed best, with 55% of matched samples falling within the expected error (within EE), higher than the 53% of VIIRS. However, the MERRA-2 product performed worst, with only 48% of matched samples falling within the EE. Based on the accuracy comparison of the same number of matched samples, MODIS DB product still performed best, with R = 0.88, RMSE = 0.17, MAE = 011 and 64% of matched samples falling within the EE. Moreover, snow cover and the complex surface features in South-West areas may limit the performance of DB algorithm products, but MODIS DB products are available in most other areas of China. Land cover type applicability analysis demonstrated that DB algorithm products performed well in low vegetation cover and impervious surface types. In the analysis of elevation applicability, the MERRA-2 product has obviously higher robustness at high elevations than the MODIS and VIIRS products. These evaluation results are expected to guide users of aerosol products in China. Highlights: AERDB captured higher AOD values in desert regions than the other products. MODIS performed best, with 55% of matched samples falling within the expected error. snow cover and the complex surface features in South-West areas may limit the performance of DB products. DB products performed well in low vegetation cover and impervious surface.